ISO 16820:2019

Overview

ISO 16820:2019 - Sensory analysis - Methodology - Sequential analysis - specifies a statistical procedure for analysing data from forced‑choice sensory discrimination tests (triangle, duo‑trio, 3‑AFC, 2‑AFC). The standard defines a sequential testing approach in which the analyst can, after each trial, decide to stop and declare a perceptible difference, stop and declare no difference, or continue testing. The sequential method often reaches a valid decision with fewer trials than fixed‑sample methods, helping to save time and resources in sensory testing.

Key topics and technical requirements

• Test types covered: triangle, duo‑trio, 3‑AFC, 2‑AFC forced‑choice discrimination tests.
• Decision framework: choose risk parameters α (alpha-risk) and β (beta-risk), and the probabilities p0 (chance performance when no perceptible difference exists) and p1 (probability of correct response when a difference exists).
• Boundary construction: the standard provides formulae and procedures to compute upper and lower decision boundaries (plotted as lines on a graph of cumulative correct responses vs. number of trials).
• Sequential stopping rules:
  • If cumulative correct responses rise above the upper line → stop and declare a perceptible difference (controlled by α).
  • If cumulative correct responses fall below the lower line → stop and declare no meaningful difference (controlled by β).
  • If cumulative results fall between the lines → continue testing.
• Statistical conversions: guidance on using the Thurstonian δ approach to convert sensory effect magnitude to the appropriate p1 value is included (Reference [7] cited).
• Annex with worked examples: practical examples (trainee selection, quality control of stored beef patties) illustrate constructing graphs and interpreting results.

Applications and who uses it

ISO 16820:2019 is intended for:

• Sensory scientists and laboratories conducting discrimination testing.
• Food and beverage R&D teams assessing formulation, ingredient or process changes.
• Quality assurance and production for monitoring product consistency, shelf life, packaging or storage effects.
• Panel managers for selecting, training and monitoring assessors (individual or panel performance can be tracked sequentially).

Practical benefits:

• Faster decisions with fewer samples and less assessor fatigue.
• Clear, risk‑controlled framework for pass/fail decisions in QC and assessor performance evaluations.
• Applicable to both panel-level and individual assessor sequential analysis.

Related standards

• ISO 5492 - Sensory analysis - Vocabulary (normative reference used by ISO 16820).
• Prepared by ISO/TC 34 / SC 12 - Sensory analysis (context for implementation).

Keywords: ISO 16820, sensory analysis, sequential analysis, discrimination tests, triangle test, duo‑trio, AFC, sensory panels, alpha risk, beta risk, Thurstonian delta, quality control.